Despite the pandemic, the exodus of tech companies, and the feared “loop of doom,” a one-two punch of return-to-office mandates and an explosion of AI startups has pushed San Francisco’s tech centers to smaller neighborhoods.
Together AI is one of several startups based in the Design District, where office space costs about $59.53 per square foot, according to commercial real estate firm Cresa’s Q4 2025 report. That’s relatively affordable compared to the cost of commercial space, which can run more than $70 per square foot in the Financial District and Mission Bay, where OpenAI’s offices are located.
CEO Vipul Ved Prakash told Business Insider that the area’s creative atmosphere and convenient parking have led to more AI companies relocating within a one-mile radius, and many of Together AI’s customers have moved right next door to the company.
“Sometimes we debug problems together and think about partnerships and other things together,” Prakash said. “It’s easier to do it in person. Once you get a certain number of companies, I think eventually the network effect will work like this.”
AI-focused companies such as Virtual AI, Applied Compute, Rox AI, and Curai Health are all housed in the same building as Together AI. Other AI companies with a three-block radius include Scale AI, Resolve AI, and Gamma. Together AI is currently negotiating a new round of funding at a valuation of $7.5 billion. These companies have a combined valuation of more than $40 billion.
“Everything happened very naturally,” Prakash added.
The current wave of startups is not the first to flood into smaller areas like the Design District, which has traditionally been an industrial hub for material supply warehouses, furniture manufacturing and repair shops. Previous waves of technology and innovation were already showing trends in rethinking where the office should be. Airbnb, for example, opened its headquarters in the Design District in 2013 and renewed its 10-year lease in 2024.
Startup employees have different needs
Richard Florida, an urban research theorist and professor at the University of Toronto, told Business Insider that traditional office districts like the financial district were built at a time when companies were stacking knowledge workers in vertical towers, but AI startups are rapidly changing that.
“The office district is no longer an office district. It’s a mixed-use district,” Florida said. “Start-ups and interesting creative companies need cheap, reconfigurable space, space that can be added on the fly, and you can’t find that in an office tower.”
“What we’re learning in the AI space is that people are working much longer hours,” Florida added. “So people are in the office from 9 a.m. to 11 p.m., so they need things around them that keep them energized, whether it’s going for a break, going for a walk, or doing something outdoors to recharge. We can’t afford to have streets empty past 5 p.m.”
Geetha Rajan, director of strategy and business operations at Freshworks who specializes in AI strategy, told Business Insider that she believes the shift in startup culture will push significant economic activity into parts of San Francisco.
“Years ago, the big boast for companies was, ‘We need to hire 10,000, 20,000, or 50,000 more people,'” Rajan said. “But when you think about it now, it’s a 10-person startup that’s already made millions of dollars.”
“Although AI allows us to go remote, the people and companies building it still very much want to meet in person to whiteboard and brainstorm,” Rajan added. “Because at the rate these companies are going, Zoom can’t survive it.”
potential beneficiaries
Rajan said the food delivery and catering scene, as well as local coffee shops, are likely the biggest winners, as employees at startups stay in the office late and come into work on weekends with little choice but to order food into their workplaces.
For example, DoorDash notes in its annual report that weekday lunch deliveries to offices and other San Francisco commercial addresses increased 15.7% from 2024 to 2025.
Rajan also said that many of the hackathon events and meetup-style activities for construction companies are now being held in quiet bars in small neighborhoods not far from startups’ offices, bringing life to places that people may have forgotten about during the pandemic lockdown.
Between the Mission District and the Castro, several up-and-coming AI companies also have offices tucked away along the main street, including Comp AI and Ace AI, just a few blocks from AI Start Academy.
Many startups, like Together AI, are moving into previously vacant spaces in non-residential areas. Notably, former OpenAI CTO Mira Murati’s startup Thinking Machines Lab occupies a 72,000-square-foot, three-story building in Northeast Mission that was once an old factory.
“When I often go to these events at the bars in the Mission, I think, ‘Wow, I haven’t been here in a long time,'” Rajan said.
Previous startup waves were different
But some economic experts are skeptical that the AI boom will directly lead to substantial economic growth for small regions or entire cities.
Ted Egan, San Francisco’s chief economist, told Business Insider that while the opening of a new company is almost always, in theory, good for the service-based businesses right next to it, this new class of startups looks very different from previous waves.
Compared to the previous technology boom in mid-market areas, While the former Twitter brought an increase in economic activity when it moved there and brought with it major companies like Uber and Square, AI startups won’t leave the same economic footprint.
“We don’t know if we’re going to see the same kind of local economic impact that we’ve seen with past technology booms,” Egan said. “It’s not that no one is hiring, but tech job postings in the city are about half of what they were before the pandemic.”
Egan pointed to city data showing sales tax revenue within the Mission increased by 8% in 2025, but he’s not convinced that’s being driven by AI startups.
“I think there’s a cluster effect with these companies,” Egan added. “It’s just that the spending profile of companies that are getting VC these days is very different than before, because AI is very computationally intensive.”